sign in sign up
<<
Home , DNA replication, Replicon (genetics)

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector , Vol. 82, 535-542, August 25, 1995, Copyright 0 1995 by Cell Press Prokaryotic Meeting Review and Eukaryotic Replicons

Joel A. Huberman complexed with viral DNA molecules, are po- Department of Molecular and Cellular sitioned at the 3’ ends of the parental duplex. Using the Roswell Park Institute penultimate 3’ deoxyribosylthymine (dT) as template, the Buffalo, New York 14263 polymerase then catalyzes the formation of a covalent bond between deoxyribosyladenine (dA) and a serine of TP. Next, the TP-dA slides back 1 nt to position the dA adjacent to the 3’ terminal dT of each strand. The poly- The model, proposed by Jacob and Brenner merase then catalyzes processive elongation of the new (1963) to explain the regulation of prokaryotic DNA replica- -primed strands, using the 3’-ended strands as tion, has proved remarkably robust. Despite the manyvari- templates and displacing the 5’-ended strands. This “slid- ations in initiation mechanism discovered since 1963, the ing-back” initiation mechanism, which provides an oppor- replicon model may be applicable, with minor modifica- tunity during subsequent replication rounds to correct nu- tions, to regulation of DNA replication in all . cleotides incorrectly incorporated at the first step, appears For several years, the Replicon Club of Paris, a group to be a common feature of linear replicons employing the of French scientists interested in regulation of replication, TP mechanism. has been meeting to discuss new results. This year, with Rolling Circle support from the French Centre National de la Recherche S. D. Ehrlich (Institut National de la Recherche Agrono- Scientifique (CNRS), the Replicon Club sponsored an in- mique [INRA], Jouy en Josas, France) discussed the Rep ternational Jacques Monod Conference on Prokaryotic protein of pC194, a rolling circle of gram-positive and Eukaryotic Replicons. The meeting was organized by . A major difference between pC194 and the well- G. Buttin (Institut Pasteur, Paris) with assistance from A. studied rolling circle replicon @X174 is that @X174 is a Falaschi (International Centre for Genetic Engineering with runaway replication, while pC194 is a plasmid and Biotechnology, Trieste, Italy) and M. Kohiyama (Insti- with regulated replication. Thedifferencecan beexplained tut Jacques Monod, Paris) and was held at the CNRS’s by comparison of the two Rep . Both proteins initi- Centre Paul Langevin in the Alpine village of Aussois, ate replication by making nicks at specific sites in one of France, from June 18-22, 1995. There, invigorated by the DNA strands, yielding 3’ ends that serve as primers beautiful surroundings and excellent cuisine, scientists and S’ends covalently attached to a Rep protein . from around the world provided evidence both for the vital- After one round of synthesis, second nicks are made at ity of the replicon model and for substantial advances in the same sites. The @X174 Rep protein employs a second our understanding of the regulation of replication. tyrosine to catalyze the second , permitting retention of the5’end and, ultimately, reinitiation. In contrast, PC194 Variety of Replicons appears to use glutamic acid-catalyzed hydrolysis for the The original replicon model (Figure 1) suggested that repli- second nick, thus losing bond energy and preventing reini- cation is positively controlled by an initiator protein that tiation. acts on a single replicator to initiate replication of a circular Rolling circle replication is also employed by geminivi- replicon. Extension of the model to other organisms re- ruses, whose small circular single-stranded rep- quires allowance for multiple linear in eu- licate via double-stranded intermediates in the nuclei of karyotic cells, each possibly having multiple replicators cells. Since replication is catalyzed entirely by en- (Figure 1). This extension requires that replicon be rede- zymes from the host cell (except the viral Rep protein) and fined to mean the stretch of DNA replicated from a single viral DNA is packed into , study of geminivirus replicator. An updated model also requires allowance for replication appears to be a promising way to learn more the possible existence of multiple initiator proteins, which about DNA replication and the in . B. Gro- may form a complex (as in Figure 1) or may separately nenborn (Institut des Sciences V&g&ales, Gif sur Yvette, bind to different portions of the replicator, and recognition France) reported that geminivirus Rep proteins nick the that interactions between initiators and replicators are typi- origin via a tyrosine and have a DNA-independent ATPasel cal binding reactions, governed by the laws of mass action GTPase activity that is essential for in vivo replication and (Figure 1). Consequently, if the concentration of initiators reminiscent of the GTPase in signal G pro- is high or their specificity is low, most or all DNA sequences teins. C. Gutierrez (Centro de Biologia Molecular Sever0 may be able to serve as replicators. Ochoa, Madrid) showed that geminivirus Rep contains an At the meeting in Aussois, developments relating to a LXCXE motif capable of mediating stable binding to the wide variety of replicons were presented. human retinoblastoma (Rb) protein. This motif is essential Protein Priming for in vivo replication, suggesting that viral replication is The phage, @29, has a linear double- linked to cell cycle control and that plants, like , stranded DNA with a viral terminal protein (TP) covalently may contain Rb homologs that inhibit entry into . attached to its 5’ ends. L. Blanc0 (Centro de Biologia Mo- Unidirectional 8 lecular Sever0 Ochoa, Madrid) described the mechanism C. Bruand and E. Le Chatelier (INRA, Jouy en Josas, of initiation of @29 replication. For each round, new TPs, France) described the pAMP1 replicon, a plasmid of gram- Cell 536

hr after infection. Evidence for interactions between UL9 and cellular DNA polymerase a, but not the viral DNA polymerase, suggests that initiation at viral origins may be accomplished by UL9 and cellular . Later Original Model rolling circle replication is independent of UL9 but requires the other viral replication proteins, which, according to physical and electron microscopic evidence, may exist as a complex. Gel shift and footprinting experiments described by P. Elias (Gdteborg University, Sweden) suggest that a protein complex may also function at the viral origin. The complex would consist of two dimers of UL9, interacting through their C-terminal domains both with specific binding sites in the origin and with four viral single-stranded DNA-binding proteins, which are called ICP-8. Since ICP-8 binds more Replicator Initiator Replicator Replicator Initiator tightly to single-stranded DNA than to UL9, it is likely that Gene1 Gene2 once single-stranded DNA is exposed at the origin by un- Figure 1. Past and Present Replicon Models winding of parental strands, ICP-8 is transferred to the (Top) Model proposed by Jacob and Brenner (1963). (Bottom) Model single-stranded DNA. arising from discussions at the Aussois meeting described in this re- During the early stage of h replication, view. initiation is mediated by the phage-encoded 0 protein at orik. Under in vivo conditions, near orid, usu- ally from the PR promoter, is also required to initiate replica- positive bacteria, which bears some resemblance to the tion. G. Wegrzyn (University of Gdansk, Poland)discussed well-known ColEl replicon. For pAM81, a plasmid-encoded circular molecules derived from h but missing most of the positive regulator, the RepE product, is essential for hgenome except the origin region. He presented evidence replication, and DNA polymerase I is replaced by DNA that 0 protein, along with other proteins involved in initia- polymerase Ill holoenzyme (Hpollll) about 200 bp down- tion, is retained at one of the two daughter origins during stream of ori. According to M.-A. Petit and L. Janniere each replication round. Such inherited replication com- (INRA, Jouy en Josas, France), this polymerase switch is plexes may mediate further replication rounds. Runaway aided by the pAM81 resolvase, which binds tightly about replication is prevented by the requirement that each 250 bp downstream of ori, thereby blocking polymerase round be activated by transcription at PR. Such transcrip- I and creating a D loop with an exposed assem- tion is mediated by the host dnaA protein, which is present bly site where an Hpollll-based replication fork can effi- in high concentration during only a limited portion of each ciently be set up. V. Bidnenko (INRA, Jouy en Josas, cell cycle. France) showed that the plasmid-encoded , Bidirectional 8 topp, also assists by relaxing plasmid DNA and thereby The single initiator protein of , encoded removing thedriving force for strand unwinding when poly- by the dnaA gene, governs initiation at the single repli- merase I reaches about 190 bp downstream of ori. The cator, oriC (reviewed by Kornberg and Baker, 1991). Inter- presenters proposed that similar mechanisms may facil- actions between the dnaA protein and oriC were described itate polymerase switching in ColEl and eukaryotic rep- by W. Messer (Max-Planck-lnstitutfur Molekulare Genetik, I/cons. Berlin). Within 0% there are four 9 nt repeats, called dnaA Previously studied prokaryotic and eukaryotic circular boxes, that serve as specific binding sites for the dnaA containing two replication origins utilize only one protein. Measurements of binding to the individual boxes origin at a time. J. B. Schvartzman (Centro de Investigaci- and tocomplete o&revealed that binding is highly context ones Biologicas, Madrid) reported that occasionally mole- dependent. The dnaA protein does not contain any of the cules of the plasmid pPl21, containing two unidirectional known DNA-binding motifs. Protein fusions and muta- ColEl origins in opposite orientation, employ both origins tional analysis revealed that the C-terminal 94 amino at once, resulting in stable “bubble” structures detectable acids, including three a helices, are responsible for DNA by two-dimensional gel electrophoresis and electron mi- binding. croscopy. Initiation of SV40 DNA replication requires binding of Bidirectional 0 Early, Rolling Circle Later two back-to-back hexameric complexes of the SV40 initia- R. Skaliter(Stanford University, California) described simi- torlhelicase T antigen to the SV40 replication origin. E. larities between herpes virus replication and a replication. Fanning (Vanderbilt University, Nashville, Tennessee) de- In both cases, early origin-dependent &mode replication scribed the importance of phosphorylation at Thr-124 of is followed by subsequent rolling circle replication to gen- T antigen for cooperative interactions between the two erate that are packaged into viral particles. hexamers, which are essential for processive bidirectional In the case of herpesvirus, two-dimensional gel electro- DNA unwinding from the origin. Similar interactions be- phoretic evidence suggests origin-specific initiation medi- tween may be important for unwinding at cellular ated by the viral origin-binding protein UL9 as early as 2 origins. Meeting Review 537

In vivo, both of the bovine papilloma virus (BPV) El and rich than S. cerevisiae and S. pombe ARS elements, and E2 proteins are essential for initiation of replication. Like do not support extrachromosomal replication of plasmids SV40 T antigen and herpes UL9, El is an origin-binding lacking . protein with activity. E2 is a transcription factor The slime mold Physarum polycephalum can exist as capable of forming a complex with El. P. Clertant (Univer- a giant single cell containing millions of naturally synchro- sity of Nice, France) reported that his laboratory has devel- nous nuclei. G. Pierron (CNRS, Villejuif, France) took ad- oped an origin- and El-dependent, but EZindependent, vantage of this synchrony to demonstrate that replication BPV replication system asefficient as that for SV40. initiates in or near the promoters of four that repli- In vitro E2 appears to only inhibit nonspecific initiation, cate in early S phase, suggesting the presence of specific but invivo it may assist in “opening”chromatin at the origin. replicators. M. Benard’s (CNRS, Villejuif, France) demon- Since the BPV in vitro system catalyzes multiple replica- stration that allelic origins are activated simultaneously tion rounds, additional factors, possibly of cellular origin, also suggests specific replicators. may be responsible for limiting BPV replication in vivo to an average of one round per cell cycle. To What Extent Does the Replicon Model Apply Epstein-Barr virus (EBV) can establish a latent infection to Ceil DNA Replication? of human B lymphocytes in which each large circular viral For the past 5 years, investigators of DNA replication in DNA molecule is replicated once per cell cycle. Previous animal cells have been facing a paradox. Some results investigations employing plasmids containing small frag- indicate that replication initiates at specific locations, and ments of the EBV had suggested that EBV latent other observations suggest that replication initiates ran- replication depends on an origin sequence, oriP, a virally domly within broad initiation zones (reviewed by Coverley encoded protein (EBNAl), and cellular replication proteins and Laskey, 1994). Similarly conflicting results were pre- (reviewed by Yates, 1993). R. Little (Albert Einstein Col- sented in Aussois. A. Falaschi (International Centre for lege of , New York) reported that replication of Genetic Engineering and Biotechnology, Trieste, Italy) de- large EBV genomes initiates both at oriP and also in scribed the presence of short nascent strands, indicative broader zones located elsewhere in the EBV genome. This of a replication origin, in all tested human cell lines within combination of specific initiation and initiation in broader an -500 bp region at the 3’ end of the human B2 zones is similar to initiation in mammalian chromosomes gene. K.-l. Tsutsumi (Iwate University, Ueda, Japan) ob- (see below). served that an - 900 bp fragment encompassing the pro- Bidirectional Linear moter of the rat aldolase B gene is capable of promoting Replication of chromosomes autonomous replication in cultured cells. M. Zannis- is largely consistent with the current replicon model (Fig- Hadjopoulos (McGill University, Montreal) reviewed stud- ure 1). Binding of an initiator protein complex (ORC) to a ies from her lab that indicate, by several methods, that replicator (autonomously replicating sequence [ARS] ele- many of the short nascent strands generated in earliest ment) is essential for initiation at or near the ARS element S phase are associated with chromosomal replication ori- (reviewed by Newlon and Theis, 1993). ARS elements are gins and are themselves capable of stimulating replication identified by their ability to serve as replication origins in in vivo and in vitro (see below). G. Wahl (Salk Institute, La plasmids. Interestingly, some ARS elements do not serve Jolla, California) provided evidence, obtained with several as origins in the . M. Weinberger (Roswell techniques, for initiation within discrete loci near the hu- Park Cancer Institute, Buffalo, New York) reported that man j3-globin gene (see below) and, interestingly, within one such chromosomally inactive ARS element, ARS307, the Syrian hamster CAD gene. The CAD gene is tran- contains the sequence motifs found in typical chromosom- scribed in early S phase, overlapping the time at which it ally active ARS elements: an 11 bp ARS consensus se- is replicated. quence and an essential flanking sequence. J. Diffley (lm- Several other investigators described evidence for perial Cancer Research Fund, Clare Hall Laboratories, broad initiation zones. P. Dijkwel (University of Virginia, South Mimms, England) added that ARS307, in its normal Charlottesville) detected a broad initiation zone encom- location on chromosome Ill, yields prereplicative and post- passing the transcriptionally silent rhodopsin gene in Chi- replicative footprints (see below) that are similar to those nese hamster ovary (CHO) cells. J. Hamlin (University of of chromosomallyactive ARS elements. Thus, the reasons Virginia, Charlottesville) found that initiation downstream forthechromosomal inactivityofARS307 remain obscure. of the dihydrofolate reductase (DHFR) gene in CHO cells In the Schizosaccharomyces pombe and Yar- is distributed over a broad zone regardless of whether rowia lipolytica, chromosomal DNA replication also is de- the region is amplified. Despite the presence of multiple pendent on and initiates at or near defined sequence potential initiation sites, the zone as awhole isnot efficient elements, as described, respectively, by J. Huberman but is frequently replicated passively by forks entering (Roswell Park Cancer Institute, Buffalo, New York) and P. from flanking regions. R. Little (Albert Einstein College of Fournier and L. Vernis (INRA, Thiverval-Grignon, France). Medicine, New York) reviewed his observation that repli- In S. pombe, the number of sequence motifs contributing cation initiates at multiple sites within the nontranscribed to ARS function (two essential motifs and more than eight spacer of human rDNA, and M. Debatisse-Buttin (Institut stimulatory motifs in the case of ars3002) is larger than Pasteur, Paris) described evidence for initiation within in S. cerevisiae. In Y. lipolytica, the origins contain no broad nontranscribed stretches in a multigenic region near obvious consensus sequence, are considerably more GC the mammalian AMPDP gene. M. Calos (Stanford Univer- Cell 533

Early Embryo ity. Potential initiation sequences would have to be present at a frequency of one or more per kilobase, but most of them would be repressed by chromatin structure. Tran- scription is a major factor in determining chromatin struc- Late Embryo ture and therefore would play a key role in specifying which and Adult sites would be used.

Figure 2. The Transition from Random to Nonrandom initiation of There was general agreement that further progress in Replication Coincides with the Beginning of Embryonic Transcription understanding animal cell replication origins requires ge- during Early Embryogenesis netic experiments to identify precisely the &-acting se- The diagram summarizes results presented by 0. Hyrien (Institut quences responsible for initiation sites and initiation Jacques Monod, Paris) concerning initiation of replication in zones. Once these sequences are identified, it should be rDNA, but the general concept may be applicable to the rest of the genome. The small stippled boxes represent initiation sites, the large possible to identify interacting proteins and determine open boxes are genes, and the horizontal arrows represent transcripts. whether those proteins serve as true initiators, serve to establish transcriptional patterns, or function in some other way. sity, California) provided evidence for sequence-indepen- dent plasmid replication in human and Drosophila cells. Genetic Experiments Suggest Multiple Determinants Attempts to rationalize these apparently contradictory of Mammalian Replication Origins observations were the subject of extensive discussions. The results of two recent genetic experiments on mamma- All participants agreed that the evidence for sequence- lian origins were described at the meeting. These and pre- independent replicationduringearlyXenopusembryogen- viously reported experiments are summarized in Figure 3. esis is convincing (reviewed by Coverley and Laskey, An initiation zone of more than 55 kb is located down- 1994) and that initiation is not completely random in adult stream of the DHFR gene in CHO cells. Within this zone, cells (it occurs at specific sites or in broad initiation zones, regions called ori8 and oriy initiate replication at relatively but not everywhere). 0. Hyrien (Institut Jacques Monod, highfrequency(reviewed by Hamlin et al., 1994). J. Hamlin Paris) presented data indicating that the developmental (University of Virginia, Charlottesville) has found that a 75 transition from completely random initiation in Xenopus kb deletion upstream of the DHFR gene that includes the rDNA to initiation primarily within the nontranscribed DHFR promoter and thereby prevents DHFR transcription spacer starts at the midblastulastage, the time when rDNA leads to loss of detectable initiation within the zone. Pre- transcription begins (Figure 2). This observation suggests viously, Handeli et al. (1989) reported that when a 16 kb that the chromatin restructuring necessary to limit tran- stretch of DNA containing orip (ADGA; Figure 3A) is scription to certain regions may also serve to limit initiation transplaced to other locations in the CHO genome, or& of replication to certain regions. sequences near it, or both continue to initiate replication. How can these observations be accommodated by the Considering the new data from J. Hamlin, it is difficult to replicon model? Two nonmututally exclusive hypotheses understand the Handeli et al. (1989) result unless trans- were proposed. M. Mechali (Institut Jacques Monod, placement was favored to regions that were also favorable Paris) suggested that specificity is lost in the early embryo for replication initiation. Repetition of that experiment with because of the high concentration of maternally derived improved characterization of the locations of the transplaced initiation proteins. Reduction of initiator concentration and hD6A segments would be useful. establishment of transcriptionally active chromatin (Figure Both Kitsberg et al. (1993) and G. Wahl (Salk Institute, 2) at the midblastula transition may explain the appear- La Jolla, California) have obtained evidence for a narrow ance of initiation zones. M. Calos (Stanford University, initiation zone or specific initiation site near the 5’ end of California) proposed that there are no rigorous sequence the human 8-globin gene (Figure 38, normal). Kitsberg requirements for initiation even in adult cells. If an initiator et al. (1993) found that a deletion covering the preferred is required, that initiator must have low sequence specific- initiation site eliminated initiation in the region, leading to

A DHFR Region Figure 3. Effects of Deletions and a Trans- placement on Mammalian Origin Function I 100 kb The open arrows represent genes, and the open circles represent sites or regions of pre- I Deletion I , hD6A , ferred initiation. The long, thin arrows in (B) J. H&in. Aussois, 1995 Handel1 et al. (1989) represent preferred directions of replication fork movement. LCR, locus control region. B 8-Globin Region

- b Kitsberg et al. (1993) Deletion I Deletion I A 0. Wahl. Aussois, 1995 Meeting Review 539

replication of the whole P-globin domain from left to right Regulation of Initiation of Replication (Figure 3B). G. Wahi and collaborators have now observed Methylation that a 30 kb deletion, which includes the locus control Both bacteria and higher organisms permit only a single region (the LCR, essential for regulation of developmen- initiation per replicon per cell cycle. One of the mecha- tally timed from the P-globin domain), nisms contributing to this limitation in E. coli is prevention eliminates initiation nearthe fl-globin gene, but in this case of premature reinitiation by temporary membrane seques- the domain is replicated from right to left. tration of newly replicated oriC. Newly replicated oriC can These four experiments suggest that the sites where be recognized by its unique methylation state: for a short DNA replication initiates in mammalian chromosomes can time after replication, parental strands are methylated, but be specified both by local sequences and by sequences daughter strands are unmethylated. E. Boye (Institute for distant from the initiation site(s). Higher resolution genetic Cancer Research, Oslo) described the protein SeqA, experiments are now needed to determine whether these which binds tightly to hemimethylated double-stranded sequences are initiator-binding sites or elements control- DNA and is essential for sequestration of newly replicated ling transcription or chromatin structure. Whatever the an- oriC. Interestingly, SeqA does not bind at all to nonmethyl- swers, the apparent requirement for distal sequences by ated , binds without sequence preference to hemi- two (out of two tested) mammalian origins suggests a type methylated DNAs, and specifically binds oriC in the fully of origin regulation in mammalian cells that has not been methylated state. Specific binding to fully methylated oriC found in cells. is about 1 O-fold weaker than nonspecific binding to hemi- methylated 0%. The ability of crude membrane fractions to bind or/C is fully accounted for by the SeqA present in those fractions. Effects of Chromatin Structure and Transcription on Initiation of Replication Cell Cycle-Dependent Kinases Regulation of initiation of replication by chromatin struc- Several years ago, Virshup (1990) suggested that the mechanism limiting initiation of eukaryotic replication to ture and transcription is not unique to eukaryotic cells. J. RouviBre-Yaniv (Institut de Biologie Physico-Chimique, a single event per replicon per S phase might be related Paris) reported that the E. coli -like protein HU, to cell cycle-dependent phosphorylation. The essence of his model was the concept that the proteins required for which can affect the bending and supercoiling of DNA, initiation can exist in multiple cell cycle-specific phosphor- can also modulate the binding of other proteins to oriC. Nearby transcription activates both oriC (reviewed by ylation states, with initiation requiring an ordered progres- Kornberg and Baker, 1991) and o& (G. Wegrzyn, Univer- sion through the different states. Within the last few years, clues regarding the mechanism(s) responsible for limiting sity of Gdansk, Poland), apparently by altering local DNA structure (reviewed by Kornberg and Baker, 1991). For initiation have appeared with accelerating frequency. It now seems that multiple redundant mechanisms may be several replicons, like ColEl and possibly pAMP1 (C. responsible. Although the picture is still far from complete, Bruand, INRA, Jouy en Josas, France), transcription is essential to generate primer . it appears possible that all of these mechanisms may be Relationships between transcription and replication are regulated by ordered cell cycle-dependent phosphoryla- tion, as proposed by Virshup (1990). abundant in eukaryotic cells. In addition to the examples The Prereplica five Complex already mentioned, D. Jackson (Oxford University) pre- J. Diffley(Imperial Cancer Research Fund, Clare Hall Lab- sented results suggesting that the mammalian replication oratories, South Mimms, England) described chromatin “factories” (large intranuclear complexes of replication en- footprinting and genetic studies of the protein complexes zymes) active in earliest S phase are located at or near the transcription factories active at the GI/S transition. at yeast ARS elements. During and most of , a postreplicative footprint is present, which ap- pears to be due to binding of ORC. From late mitosis through Gl, a broader, stronger prereplicative footprint is Replication within the Ceil Cycle evident. The protein, as well as ORC, is required In most eukaryotic cells, different replication origins fire for establishment of the prereplicative complex and for at different times during S phase. W. Fangman (University initiation of DNA replication. The DBF4 and CDC7 proteins of Washington, Seattle) described identification of several are also likely to be components of the prereplicative com- late-firing S. cerevisiae origins, some close to and some plex, at least in late Gl, and are essential for initiation. far from . Late firing of the -proximal Other proteins, such as the mainte- origins is a consequence of telomere proximity as illus- nance (MCM) proteins (see below), may also be compo- trated, for example, by the fact that transplacing an early- nents of the prereplicative complex. Formation and disso- firing origin to a position close to a telomere renders it late ciation of the prereplicative complex are clearly an ordered firing. In contrast, late firing of origins far from telomeres series of events essential for initiation. The extent to which appears to be due to chromosomal sequences flanking these events are regulated by cell cycle-specific phos- these origins, called delay elements. The se- phorylation remains to be determined. quences of delay elements do not resemble those of telo- MCM Proteins meres. Complementary evidence regarding limitation of eukary- Cell 540

otic replication comes from converging investigations of lation. Thus, the ordered, cell cycle-specific progression Xenopus and yeast MCM proteins. Licens- of the MCM proteins through their phosphorylation states ing factor was hypothesized by Blow and Laskey (1988) may contribute to licensing. to be essential for initiation of replication and able to asso- The B Subunit of DNA Polymerase a ciate with chromatin only during mitosis or if the nuclear P. Plevani (Universita degli Studi di Milano) reported that membrane were permeabilized. Biochemical assays for the second largest (B) subunit of DNA polymerase a has Xenopus licensing factor (Chong et al., 1995; Kubota et a unique role in initiation of replication. Furthermore, the al., 1995; Madine et al., 1995) have identified some of its B subunit is unphosphorylated in late mitosis through Gl components as members of the family of MCM proteins, and becomes phosphorylated near the GllS interface. proteins essential for DNA replication, originally identified The implications of these changes in phosphorylation for in S. cerevisiae but now known to be present in all eukary- B subunit function are not yet known, but the available data otic organisms (reviewed by Tye, 1994). Ft. Laskey (Well- suggest that the ordered cell cycle-specific progression of come/CRC Institute, Cambridge) described studies (Mad- the B subunit through its phosphorylation states may also ine et al., 1995) revealing that several Xenopus MCM contribute to licensing. proteins form a coimmunoprecipitable complex that is es- Replication Factories sential for initiation. Unlike the originally postulated licens- As pointed out by D. Jackson (Oxford University), replica- ing factor, Xenopus MCMs can be transported into intact tion appears to take place in factories, large complexes nuclei. It is possible that another component of biochemi- of proteins involved in DNA replication. These factories cally defined Xenopus licensing factor, the B component include DNA’ polymerase a and (Chong et al., 1995) can gain access to chromatin only (RPA), but when active do not contain MCM proteins during mitosis or if the nuclear membrane is permeabil- (based on the observations of R. Laskey, S. Kearsey, and ized. Because the biochemical studies have revealed that M. Mechali). Multiple replication forks are elongated within licensing factor consists of multiple proteins (at least four each replication factory, rendering the factories large members of the MCM family plus an unknown number of enough to be detectable by electron microscopy. Factories proteins in the B component), R. Laskey suggested that are notvisible by electron microscopy until late Gl , and the the term licensing factor is no longer useful. The word earliest factories appear adjacent to sites of transcription. licensing, however, may still be used to describe the over- During S phase, the distribution of factories within the nu- all process whereby initiation is limited to one event per cleus changes, and the factories become larger. Thus, replicon per cell cycle. replication factories, too, are regulated during the cell cy- S. Kearsey (Oxford University) pointed out that the hu- cle. The role of phosphoryation in this regulation remains man MCM2 homolog, BM28, is found in the nucleus to be elucidated. throughout the cell cycle but is tightly bound to chromatin The results presented in Aussois concerning cell cycle only during Gl phase. The tightly bound form is gradually control of replication are partially summarized in Figure converted to the loosely bound form during S phase, and 4. In G2 and early mitosis, only initiator proteins (here there is no colocalization of the tightly bound form with active DNA replication factories, consistent with observa- tionsof R. Laskeyand M. Mechali (Institut Jacques Monod, G2 Paris). T. Su (University of California, San Francisco) provided Mitosis evidence for at least seven different MCM proteins in Dro- I sophila. These proteins exist in large complexes of about 600 kDa. At least two different complexes can be distin- Late mitosis/ Early Gl guished on the basis of the MCMs they contain. These complexes are surprisingly salt stable, resisting even 2 M Assembly of replication factones NaCI. Completion of pre-replicative complexes M. Mechali observed that the Xenopus homolog of an S. pombe MCM, cdc21, associates with chromosomes in punctate fashion at a very early stage of nuclear formation End in Xenopus extracts, even earlier than proliferating cell ofG1 nuclear antigen (PCNA) association. Later the staining be- Delivery of origins to replication factories? Release of CDCG, MCMs, other proteins comes more diffuse. During S phase, Xenopus cdc21 is Phosphorylation of MCMs, Ei subunit, others Initiation of replication detected only in nuclear regions that have not yet repli- S phase cated. J M. Mechali and S. Kearsey noticed that XenOpUS cdc21 G2 and human BM28, respectively, are hypophosphorylated during Gl and hyperphosphorylated during G2 and M Figure 4. Assembly and Disassemblyof a PrereplicativeComplexdur- phase. The effects of these phosphorylation changes on ing the Cell Cycle MCM function are not yet known, but correlations between The open circles labeled A, B, C, and D represent proteins of the phosphorylation state and extent of binding to chromatin prereplicative complex that have not yet been identified. It is possible suaaest that bindina is likelv to be reaulated bv phosphorv- that some of these miaht be MCM oroteins. Meeting Review 541

represented by the S. cerevisiae ORC) are bound to ori- vant to DNA replication were provided at the Aussois meet- gins. Additional proteins, including the CDC6 protein, bind ing. For example, J. Borowiec (New York University at the end of mitosis. Genetic evidence for interactions Medical Center, New York) described experiments sug- between several of the MCMs and ORC (Loo et al., 1995) gesting that human RPA binds single-stranded DNA by suggests that the MCMs may also be included in prerepli- initially contacting a small (- 8 nt) binding site and subse- cative complexes. Additional proteins may bind at later quently reorienting to an elongated form with an -30 nt times during Gl. J. Diffley’s (imperial Cancer Research binding site. During this process, human RPA also ap- Fund, Clare Hall Laboratories, South Mimms, England) pears to undergo a significant conformational change that evidence suggests that both DBF4 and CDC7 are included can be detected by interaction with the DNA-dependent in this complex by the end of Gl. Also during Gl , perhaps protein kinase. at or perhaps independently of the prereplicative com- U. Hijbscher(Universityof Zurich) described the interac- plexes, replication factories are assembled. Initiation of tions of PCNA with DNA, with DNA polymerase 6, and replication at the Gl/S interface requires delivery of those with ~21. PCNA forms a “sliding clamp,” a molecular ring prereplicative complexes that will function earliest in S to through which DNA can slide. The accessory factor repli- the replication factories (if they are not already there) and cation factor C (RFC) can catalyze the loading of the PCNA dissociation of the now phosphorylated MCM proteins. clamp onto double-stranded DNA, but the clamp must then CDC6 and (presumably) other proteins of the prereplica- slide along the DNA to a 3’-OH primer terminus before it tive complex are also released at this time. The fact that becomes competent to interact with polymerase 6. p21 the MCM proteins are released from chromatin gradually does not inhibit PCNA’s ability to slide along DNA, slightly during S phase but are never detected in association with inhibits the loading of PCNAonto DNA, and strongly inhib- active replication factories suggests that the MCM pro- its PCNA’s association with polymerase 6. ~21’s ability to teins may help to deliver origins to replication factories (if inhibit replication selectively with minimal effect on DNA they are not already there), or, within not-yet-active factor- repair appears to be due to the tendency of polymerase ies, the MCM proteins may facilitate initiation and activa- 6 to fall off the DNA when it encounters a pause site, the tion of the factory and then dissociate. fact that pause sites are infrequent, so most short repair patches do not contain a pause site, and the inhibition by Control of Inappropriate Replication p21 of the reloading of polymerase 8 at the pause site. All living organisms have checkpoints to deal with prob- DNA synthesis in eukaryotic cells is thought to involve lems that may arise during cell cycle progression. S. S&or priming by DNA polymerase a- and then a switch (Institut de Genetique et Microbiologic, Orsay, France) de- to elongation by polymerase 6. G. Maga (University of scribed a novel checkpoint in B. subtilis in which inappro- Zurich) provided evidence that these two polymerases priately initiated DNAsynthesis is blocked at sites far away plus RFC form an isolatable, ATP-dependent trimeric com- (nearly 200 kb) from the origin. Blockage requires the plex. This appears to be a promising step toward the isola- alarmone ppGpp and replication terminator protein, a con- tion of a complete replication complex. trahelicase also essential for termination of normal chro- Additional promising steps toward isolation of complete mosomal replication at the normal termination site, terC. replication systems were described by D. Braguglia(Swiss The sites at which replication forks are blocked in the Institute of Experimental Cancer Research, Epalinges, checkpoint response display sequence similarity to terC. Switzerland) and M. Zannis-Hadjopoulos (McGill Univer- Genes proximal to the block include those important for sity, Montreal). D. Braguglia reported that he has devel- vegetative growth and sporulation. This reversible check- oped conditions under which yeast nuclear extracts initiate point may also act as a nutritional sensor before replication replication on naked DNA substrates. Although this repli- of the entire chromosome. cation is independent of origin sequences and ORC, it is In mammalian cells, the protein p21 (which is induced dependent on CDCG, polymerase a, and polymerase 6. by ~53) can block inappropriate DNA replication by two When intact yeast S phase nuclei are incubated in similar mechanisms: inhibition of cell cycle kinases and inhibition extracts, ORC-dependent incorporation of biotinylated of PCNA (a factor essential for eukaryotic rep- dUTP takes place at a few foci within each nucleus, remi- lication fork progression; see below). R. Fotedar (Institut niscent of the replication factories of mammalian cells. de Biologie Structurale, Grenoble, France) reported that The fact that most of the observed nuclear incorporation two different regions of p21 are responsible for these two is ORC dependent raises the possibility that ORC may different activities: the N-terminal portion mediates in- play a role in fork movement as well as in initiation. M. teractions with ~33~~~~ and A and E, while the Zannis-Hadjopoulos’ system employs HeLa cell nuclear C-terminal portion binds PCNA. Each of these domains and cytoplasmic extracts. Unlike the yeast system, it per- is capable independently of blocking SV40 replication in mits sequence-dependent replication of supercoiled plas- vitro or chromosomal replication when overexpressed in mids. The sequences active in this system appear to sup- vivo. port relatively abundant initiation in vivo. Therefore, although initiation in this system does not require passage through Gl phase and thus is not identical to initiation in and Complexes the cell, results obtained with this system may shed light Comprehension of any requires under- on some of the sequence and biochemical requirements standing at the biochemical level. Several examples rele- for cellular initiation. Cell 542

Closing Remarks Just a few years ago, questions such as the biochemistry of initiation in , the existence of specific origins in , and the regulation of replication in all organ- isms seemed to be unfathomable mysteries. As the meet- ing in Aussois demonstrated, tremendous progress has been made on all these fronts. Within the next few years, higher resolution genetic experiments are likely to help remove the remaining uncertainties about replication ori- gins in animal cells, and continued deployment of bio- chemical and genetic techniques should should permit continued rapid progress in understanding the compli- cated control of DNA replication in eukaryotic organisms. We can look forward to much excitement.

Acknowledgments

I want to thank the other participants in the Aussois meeting and sev- eral of my colleagues at the Roswell Park Cancer Institute for their thoughtful comments on the manuscript, without which this meeting summary would have been significantly less clear and accurate. I apol- ogize to the numerous presenters at Aussois whose interesting results could not be included in this summary owing to lack of space:

References

Blow, J. J., and Laskey, Ft. A. (1988). A role for the nuclear envelope in controlling DNA replication within the cell cycle. Nature 332, 541% 546. Chong, J. P. J., Mahbubani, H. M., Khoo, C-Y., and Blow, J. J. (1995). Purification of an Mcm-containing complex as a component of the DNA replication licensing system. Nature 375, 418421. Coverley, D., and Laskey, R. A. (1994). Regulation of eukaryotic DNA replication. Annu. Rev. Biochem. 63, 745-776. Hamlin, J. L., Mosca, P. J., and Levenson, V.V. (1994). Definingorigins of replication in mammalian cells. Biochim. Biophys. Acta 7798, 85- 111. Handeli, S., Klar, A., Meuth, M., and Cedar, H. (1989). Mapping replica- tion units in animal cells. Cell 57, 909-920. Jacob, F., and Brenner, S. (1963). Sur la regulation de la synthese du DNA chez les batteries: I’hypothese du replicon. Comp. Rend. Acad. Sci. 256, 298-300. Kitsberg, D., Selig, S., Keshet, I., and Cedar, H. (1993). Replication structure of the human 8-globin gene domain. Nature 366, 588-590. Kornberg, A., and Baker, T. A. (1991). DNA Replication, Second Edi- tion (New York: W. H. Freeman and Company). Kubota, Y., Mimura, S., Nishimoto, S.-l., Takisawa, H., and Nojima, H. (1995). Identification of the yeast MCMS-related protein as acompo- nent of Xenopus DNA replication licensing factor. Cell 87, 601-809. Loo, S., Fox, C. A., Rine, J., Kobayashi, R., Stillman, B., and Bell, S. (1995). The origin recognition complex in silencing, cell cycle progres- sion, and DNA replication. Mol. Biol. Cell 6, 741-756. Madine, M. A., Khoo, C.-Y., Mills, A. D., and Laskey, R. A. (1995). An MCM3 complex is required for cell cycle regulation of DNA replication in vertebrate cells. Nature 375, 421-424. Newlon, C. S., and Theis, J. F. (1993). The structure and function of yeast ARS elements. Curr. Opin. Genet. Dev. 3, 752-758. Tye, B.-K. (1994). The MCM2-3-5 proteins: are they replication licens- ing factors? Trends Cell Biol. 4, 160-166. Virshup, D. M. (1990). DNA replication. Curr. Opin. Cell Biol. 2, 453- 460. Yates, J. L. (1993). EBV DNA replication: how do viral genomes be- come amplified during latent infection? In The Epstein-Barr Virus and Associated Diseases, T. Tursz, J. S. Pagano, D. V. Ablashi, G. deThe, G. Lenoir, and G. R. Pearson, eds. (London: John Libbey Eurotext), pp. 267-272.